Remote conference system using multicast transmission for performing echo cancellation

[0002] The present invention relates to a remote conference system, and more particularly to a remote conference system which transmits voice data using multicast in a server client model.

[0003] In a conventional remote conference system, each of conference terminals and a conference controller are mutually connected. Each conference terminal transmits voice data to the conference controller. The conference controller synthesizes the voice data transmitted from each of the conference terminals, and transmits the synthesized voice data to each of the conference terminals. At the time of synthesizing the voice data, the conference controller performs echo canceling processing in order to prevent echo (e.g., howling) in each of the conference terminals.

[0004] The echo canceling processing is the one in which the voice data already transmitted from a certain conference terminal is canceled from voice data to be transmitted to that conference terminal. For example, assuming that voice data from a conference terminal A be “a”, voice data from a conference terminal B be “b”, and voice data from a conference terminal C be “c”, voice data (b+c) is transmitted to the conference terminal A, voice data (c+a) is transmitted to the conference terminal B, and voice data (a+b) is transmitted to the conference terminal C.

[0005] In the conventional remote conference system, since the different voice data are transmitted from the conference controller to the conference terminals, the conference controller must synthesize the voice data for each conference terminal one by one. In addition, the conference controller must transmit each of the voice data one by one (i.e., unicast transmission). Therefore, loads for processing in the conference controller increase and traffics on networks increase, resulting in an increase in data delay.

[0006] In view of the foregoing problem of the conventional system, an object of the present invention is to provide a remote conference system which averages the processing load over the whole of the system and suppresses the traffics on the network while possessing an echo canceling function.

[0007] In one preferable embodiment of the present invention, one conference terminal cancels its own voice data from a synthesized voice data which is obtained by synthesizing voice data produced by all of other conference terminals.

[0008] With the unique and unobvious structure of the present invention, the terminal voice data causing the echo is canceled from the synthesized voice data transmitted by a multicast transmission.

[0009] The above-mentioned and other objects, features and advantages of this invention will become more apparent by reference to the following detailed description of the invention taken in conjunction with the accompanying drawings in which:

[0010] FIG. 1 is a block diagram showing an entire configuration of a remote conference system of the present invention;

[0011] FIG. 2 is a block diagram showing a configuration of a conference terminal 100 in the present invention;

[0012] FIG. 3 is a block diagram showing a structure of a conference controller 200 in the present invention;

[0013] FIG. 4 is a format of a terminal packet generated by a terminal packet generator 130;

[0014] FIG. 5 is a format of a synthesized packet synthesized by a synthesizer 220;

[0015] FIG. 6 is a diagram showing a configuration of a terminal packet memory 150; and

[0016] FIG. 7 is a drawing showing a configuration of an echo canceler 170.

[0017] A remote conference system in accordance with a preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.

[0018] Referring to FIG. 1, in a remote conference system of the present invention, a plurality of conference terminals 100 as clients are coupled to a conference controller 200 as a server via a network 300. Each of the conference terminals 100 transmits the voice data as a terminal packet to the conference controller 200. The conference controller 200 synthesizes the terminal packets received from the conference terminals 100, and transmits the synthesizing results as the synthesized terminal packets to the conference terminals 100.

[0019] The network 300 preferably has a function multicasting the synthesized packets from the conference controller 200 to the conference terminals 100. The network 300 performs the multicasting using, for example, UDP (User Datagram Protocol) as the fourth layer protocol and ATM (Asynchronous Transfer Mode) as the second layer protocol. That is, the multicasting may be realized using any layer. It should be noted that since the packets to be transmitted to the conference terminals are the same, a network having no multicasting function can transmit the packets by the form of an unicasting.

[0020] Referring to FIG. 2, each of the conference terminals 100 comprises a microphone 110, an encoder 120, a terminal packet generator 130, a terminal packet transmitter 140, a terminal packet memory 150, a synthesized packet receiver 160, an echo canceler 170, a decoder 180 and a speaker 190.

[0021] The microphone 110 collects a voice sound having more than noise level, and converts the voice sound to a voice signal. The microphone 110 outputs the electronic signal to the encoder 120.

[0022] The encoder 120 encodes the voice signal from the microphone 110. As a coding system, for example, a PCM (Pulse Code Modulation) encoding system and an MPEG (Motion Picture Expert Group) audio encoding system are adopted. The PCM encoding system is performed in such manner that voice waveforms are sampled depending on a time axis and the sampled voice waveforms are quantized. The MPEG audio encoding system is performed in such manner that the voice waveforms are sampled depending on a frequency axis every sub-band and the voice waveforms are quantized. The encoder 120 outputs the encoded voice signal as the voice data, to the terminal packet generator 130.

[0023] The terminal packet generator 130 appends terminal identifiers and serial numbers to the voice data from the encoder 120 in order to generate terminal packets. Each of the terminal identifiers is the one that is for identifying the corresponding one of the conference terminals. When “m” conference terminals are present, each of the terminal identifiers is expressed as “ID_i” (i=1 to m). The serial number is a number, generated for each conference terminal, that is for identifying the voice data in each conference terminal. When n-series voice data are present, the serial numbers are expressed as “SEQ_j” (j=1 to n). Each of the voice data “DT_ij” is identified in the system by the corresponding one of the terminal identifiers and serial numbers. Specifically, the terminal packets generated by the terminal packet generator 130 are illustrated as FIG. 4.

[0024] The terminal packet transmitter 140 transmits the terminal packet generated by the terminal packet generator 130 to the network 300 according to the specified procedures.

[0025] The terminal packet memory 150 stores the terminal packets generated by the terminal packet generator 130. It should be noted that since each of the terminal packets stored in the terminal packet memory 150 is the one generated in the corresponding one of the conference terminals, the terminal identifiers in the terminal packets need not be stored therein. Therefore, the terminal packet memory 150 has configuration as in FIG. 6.

[0026] The synthesized packet receiver 160 receives the synthesized packet from the conference controller 200 through the network 300. The synthesized packet receiver 160 outputs the synthesized packet to the echo canceler 170.

[0027] The echo canceler 170 cancels (e.g., erases) the terminal voice data in the terminal packets stored in the terminal packet memory 150 from the synthesized voice data in the synthesized packets which has been received by the synthesized packet receiver 160. The echo canceler 170 outputs a data after the cancellation as a conference voice data.

[0028] The decoder 180 decodes the conference voice data from the echo canceler 170, and outputs the decoded conference voice data as a conference voice signal to the speaker 190.

[0029] The speaker 190 sounds the conference voice signal from the decoder 180.

[0030] Referring to FIG. 3, the conference controller 200 comprises a terminal packet receiver 210, a synthesizer 220 and a synthesized packet transmitter 230.

[0031] The terminal packet receiver 210 receives the terminal packet from each of the conference terminals 100, and outputs it to the synthesizer 220.

[0032] The synthesizer 220 synthesizes the terminal packet illustrated in FIG. 4, which is sent from the terminal packet receiver 210, and generates the synthesized packet. Synthesizing processing will be described later in detail.

[0033] The synthesized packet transmitter 230 transmits the synthesized packet from the synthesizer 220 to the network 300.

[0034] Next, the synthesizing processing for the voice data performed in the present invention will be described below.

[0035] Referring to FIG. 4, the j-th voice data DT_ijin the conference terminal i forms the terminal packet together with the identifier ID_i, and the serial number SEQ_j. Here, the voice data arrived to the conference controller 200 from the conference terminal i at a certain time T_kshall be expressed as D_ik. The conference controller 200 generates the synthesized voice data DTM_kin the following manner. ${\mathrm{DTM}}_k=\sum _{i=1}^m{D}_{\mathrm{ik}}$

[0036] For example, assuming that at the time T_k, the terminal voice data DT₁₄of the serial number 4 from the conference terminal 1 be received, the terminal voice data DT₂₃of the serial number 3 from the conference terminal 2 be received, and the terminal voice data DT₃₃of the serial number 3 form the conference terminal 3 be received, DTM_kis generated according to the following formula. $\begin{array}{c}{\mathrm{DTM}}_k=D_{1k}+D_{2k}+D_{3k}\\ ={\mathrm{DT}}_{14}+{\mathrm{DT}}_{23}+{\mathrm{DT}}_{33}\end{array}$

[0037] Referring to FIG. 5, the voice data DTM_ksynthesized in the above-described manner forms the synthesized packet together with a pair of the terminal identifier and serial number used in the synthesizing processing. This synthesized packet is transmitted to each of the conference terminals 100.

[0038] Next, echo canceling processing performed in the present invention will be described below.

[0039] For example, the voice data a, b and c are transmitted from the three conference terminals A, B and C, respectively, and they are generated as the synthesized voice data (a+b+c) by means of the conference controller 200. The synthesized data (a+b+c) is distributed to the three conference terminals A, B and C. The conference terminal A performs the subtraction processing “(a+b+c)−a” in which the voice data a is subtracted from the received synthesized voice data (a+b+c), the voice data a being transmitted by the conference terminal A. Then, the conference terminal A reproduces the voice data (b+c). The conference terminal B performs the subtraction processing “(a+b+c)−b” in which the voice data b is subtracted from the received voice data (a+b+c), the voice data b being transmitted by the conference terminal B. Then, the conference terminal B reproduces the voice data (a+c). The conference terminal C performs the subtraction processing “(a+b+c)−c” in which the voice data c is subtracted from the received voice data (a+b+c), the voice data c being transmitted by the conference terminal C. Then, the conference terminal C sounds the voice data (a+b). Thus, the echo to be caused by the voice signal of the local conference terminal can be prevented.

[0040] When the echo canceler 170 receives the synthesized packet from the synthesized packet receiver 160, the echo canceler 170 extracts the serial number corresponding to its terminal identifier from the synthesized packet. For example, in the example shown in FIG. 5, the conference terminal 100 corresponding to the terminal identifier “ID₂” extracts “SEQ₃” as the serial number.

[0041] Referring FIG. 7, the echo canceler 170 searches the terminal packet memory 150 using the serial number SEQ_xwhich has been extracted in the above-described manner. That is, the echo canceler 170 searches an entry having the serial number SEQ_x, in the terminal packet memory 150 by using a comparator 171. If the entry having the serial number SEQ_xis present in the terminal packet memory 150, a selector 172 selects terminal voice data DT_ixcorresponding to the serial number SEQ_x, of this entry. A subtractor 173 subtracts the voice terminal data DT_ixfrom the synthesized voice data DTM_k, and outputs the voice data DM_k. If no entry having the serial number SEQ_xis present in the terminal packet memory 150, processing for canceling the echo is unnecessary. Therefore, the selector 172 outputs “0”, and the subtractor 173 outputs the synthesized voice data DTM_kas the voice data DM_kdirectly. After this voice data DM_kis decoded by the decoder 180 and sounded by the speaker 190.

[0042] As is apparent from the above description, according to the present invention, since the terminal voice data by the conference terminal is stored in this conference terminal, the terminal voice data causing the echo is canceled from the synthesized voice data so that an occurrence of the echo can be prevented. Moreover, since the packet to be transmitted to each of the conference terminals is identical, higher speed processing can be achieved in the network having a multicasting function.

[0043] While the invention has been described in terms of several preferred embodiments, those skilled in the art will recognize that the invention can be practiced with modification within the spirit and scope of the appended claims.